COMPLIANCE TO GROWTH HORMONE THERAPY IN CHILDREN AND ADOLESCENTS WITH GROWTH HORMONE DEFICIENCY AND TURNER SYNDROME AND IMPACT ON HEIGHT VELOCITY: A PROSPECTIVE STUDY

Adherence to recombinant growth hormone (rGH) in children is necessary to ensure good treatment outcome. Reported non-compliance to rGH in children varied from 5-82% with little data involving Asian population. The objectives of this study are to evaluate the compliance of children and adolescents to rGH, effect of non-compliance on growth and factors affecting non-adherence. Methods: This is a prospective cohort study over 1-year aiming at all patients age less than 18 years old treated with rGH in our centre. Compliance was assessed from the number of returned medications and electronic record within device. Poor compliance was defined as utilised dose <86% prescribed (equivalent to missing injection ≥1 per week). Result: Thirty-four patients were recruited including 20 (59%) patients with GH deficiency and 14 (41%) with Turner syndrome. Poor compliance was noted in 35% of patients. Poor compliance in GHD patients was significantly associated with an older age (mean 12.55 ± 3.33 vs 9.28 ± 3.20 years old, p=0.038) and longer duration of treatment (mean 5.41 ± 3.0 vs 2.93 ± 2.18 years, p=0.046). Frequent reasons for missing doses were forgetfulness and inadequate medications. Participants with poor compliance had significantly reduced height velocity (HV) and HV standard deviation score (SDS) compared to those with good compliance (p<0.05). Conclusion: Poor compliance is more common in adolescents, those on longer duration of treatment and leads to significantly reduced height velocity. Regular monitoring throughout the course of rGH is important. Measures to improve compliance must address the underlying reasons.


Introduction
Recombinant growth hormone (rGH) had been used since 1980s [1] and had been widely used to treat various growth disorders, both growth hormone deficient (GHD) and non-GHD. However, due to its high cost, there is a stringent criterion for the usage of rGH. The Malaysian Ministry of Health Clinical Practice Guideline 2010 [2] recommends the use of GH for only three main indications; proven growth hormone deficiency (GHD), Turner Syndrome (TS), and small for gestational age (SGA) with poor catch-up growth. We are a public-funded major paediatric endocrine referral centre in Malaysia, and our pharmacy's drug expenditure report in the year 2020 showed that the expenses on rGH were more than RM550,000 per year and likely to further increase over time.
Many factors influence the outcome of rGH in children. These include the underlying diagnosis, age of treatment, and appropriate dosing. Adherence to rGH treatment in children is crucial to ensure an optimal final height outcome [3][4][5][6]. There are not many studies involving Asian children and none in our local setting. Given the high expenses of GH therapy, studies are warranted to evaluate the compliance of patients to rGH and ascertain the factors affecting adherence and suboptimal outcome [7]. In 2011, HL Ooi et al [8] evaluated the response and factors affecting the final height of 17 Malaysian children with GHD treated with rGH. The majority of them achieved final height at the lower target height range with -0.7 SDS from the mid parental height, however compliance factors were not explored.
Based on the systematic review by Fisher et al. [9], estimates of non-compliance to GH treatment in children and adolescents varied from 5 to 82%, depending on the studied population, methods and definitions used for non-compliance. We decided to adopt the approach from Cutfield et al. study [10] in New Zealand, which measure compliance by counting the returned empty vials of medication; and compliance was defined as ≥85% adherence (equivalent to missing injection <1 per week).
The objectives of our study are to evaluate the compliance of children and adolescents to rGH, the effect of poor compliance on height velocity (HV), and factors affecting adherence.

Methods
This is a prospective cohort study over a period of 1 year from the year 2019 to 2020, aiming at all patients on rGH less than 18 years old in our centre. The study had been approved by the Medical Research and Ethics Committee (MREC), Ministry of Health, Malaysia (NMRR ID: NMRR-18-2969-44525), and informed consent had been obtained from all the participants.
All recruited patients and caretakers were reviewed and interviewed thrice by the pharmacists (co-authors) using a listed questionnaire (Appendix 1) to assess factors contributing to poor compliance, injection techniques, transport, and storage of medications. This was done at recruitment and subsequent two clinic follow-ups with a 3-4-months interval between visits. Corrective measures were taken if patients were noted to have wrong injection techniques or inappropriate transport and storage of medications. During every clinic visit, height was measured using a Harpenden Stadiometer and the average height of three measurements was used. The methodology is summarized in Figure 1. * (%) = 100 (*Calculated during each visit to the hospital pharmacy upon returning used medication and collection of new ones).
Poor Compliance = utilised doses <86% of the prescribed doses @ equivalent to missed ≥1 injection a week The percentage of compliance for each patient was calculated based on the doses utilized over the doses prescribed ( Figure 1). Doses utilized were checked and calculated by the pharmacists during each visit to the pharmacy for the new collection of medications. Participants were required to bring the used medications and their GH devices during each visit. There were two types of GH device used, i.e., a single used disposable pen which is the self-inject device; and an auto-injector device equipped with an electronic device (Easypod) which auto-records the daily administration of medication and its dose. For the disposable pen, manual calculation of the utilized doses was based on the remaining doses in the returned used pen. Visits to the hospital pharmacy would coincide with the clinical review in the Paediatric Endocrine Clinic. However, there are additional visits required for the collection of medications between 1-3 monthly when there is a lack of supply in the hospital.
Poor compliance was defined as utilized doses <86% of the prescribed doses which is equivalent to missing less than 1 injection a week (one injection per week =14%) ( Figure 1). This was similar to the definition used by Cutfield et al. study [10] in New Zealand, who defined compliance as ≥85% adherence (equivalent to no more than one missed dose a week on average to prescribed treatment).

Statistical analysis
Sample size calculation was done using Sample Size Calculator for Estimations v 1.0.03 (Naing L, Winn T and Rusli BN). The estimated prevalence of non-compliance to GH therapy was 66%, from Cutfield et al. [11]. The calculated sample size for a 95% level of confidence and precision of 0.05 was 345 samples. However, as our patient pool is only around 35 patients, finite population correction was applied and the corrected sample size was 33 patients. Given the very small patient population, the decision was made to include all patients to maximize data collection.
Statistics analysis was done using SPSS version 22. Descriptive data were expressed as mean ±standard deviation (SD) unless otherwise stated.
Paired T-test was used to analyze outcomes before and after the intervention. For univariate analyses, the independent samples test was used for the analysis of normally distributed variables, while Mann Whitney Test was used for non-normally distributed data. Categorical data were analyzed using Chi-square or Fisher's exact test. For multivariate analysis, binary logistic regression was used. A value of P < 0.05 is considered statistically significant. Height velocity standard deviation score (HV SDS) was calculated using the software Auxology version 1.0 b17 Copyright® 2003 Pfizer.

Compliance rates and associated factors
The majority of the participants, 22 (65%) were adherent to treatment but 12 (35%) had poor compliance. There was poorer compliance among GHD participants as compared to TS (45% vs 21%).
Paired T-test did not demonstrate an improvement in compliance for both GHD and TS groups from the first to final assessment during the study. Among the GHD participants, poor compliance was more common in the older age group or the adolescents (mean age 12.55 ± 3.33 years) as compared to the younger patients (9.28 ± 3.20 years) with a mean age difference of 3.27 years (p=0.038). There was a significant association found between compliance and duration of treatment in GHD patients. Good compliant patients had been on a shorter duration of treatment (mean 2.93 ± 2.18 years) while those with poor compliance had been on treatment for a longer duration (5.41 ± 3.0 years, difference 2.48 years, p=0.046.). There were no significant associations found in participants with Turner syndrome. These findings are summarized in Table  3. The majority (number, n=28, 82.4%) of the participants were using the self-inject pen device and the rest auto-injector easypod. All participants said they were satisfied with the GH treatment and the devices and were aware of the importance of GH therapy and compliance to medication. Eight (23.5%) of the participants had an unsatisfactory techniques of GH administration during the initial assessment, which improved by the end of the study. Most understood the mode of transport and storage of medication though 2 (5.9%) had to be corrected and counseled, despite already being on rGH for more than 2 years.
The older group of patients (>10 years old) were more likely to self-inject (p=0.068) and had less supervision on GH administration (p=0.052) though these were not statistically significant. Multivariate analysis did not demonstrate any statistically significant association between compliance with race, gender, diagnosis, education level, the person who administer medication, supervision or device used.
Demographics and other factors explored in the questionnaire are summarized in Table 4.
The association of HV and HV SDS with compliance remained significant after excluding patients on the first year of rGH (Supplementary Table 1).

0.049
HV=height velocity, HV SDS= height velocity standard deviation score Discussion rGH had been widely used for the treatment of growth disorders in children and adolescents [12], and compliance is important to ensure efficacy and the best outcome of treatment [9,13]. A recent systematic review by Graham Selina et al. looking at studies from 1985 to 2018 measuring treatment adherence in paediatric GHD population [6] reported various methods in assessing compliance to treatment and definitions of compliance. Methods measuring medications adherence included redeemed rhGH prescriptions/vials [4,10,14], self-report questionnaires to the patient and/or parents 15,16], and electronic monitoring device in conjunction with a clinical kit software [17][18][19][20][21].
We had chosen to adopt the approach of Cutfield et al [10]. However, given the relatively small sample size, we had divided to only two categories i.e., good and poor compliance. The poor compliance rate was defined as <86% of the prescribed doses which is also equivalent to missing more than 1 injection a week (1 injection per week =14%). Good compliance was defined as ≥86% of prescribed doses which is equivalent to missing less than 1 injection a week. Using this threshold, Cutfield et al. had reported a reduced HV in their participants. Hence, we felt it was a reasonable threshold to adopt. However, we acknowledge there are variable definitions for compliance rate with different studies and this is a limitation of our study.
Poor compliance was noted in 35% of participants in our study. The compliance rate was better compared to Cutfield et al. who reported noncompliance to GH treatment in 66% of their subjects [8]. Kapoor et.al (2007) (n=75) reported 39% had missed more than 1 injection per week and 23% missed 2 injections per week [4]. Our compliance rate could have been higher as the participants were aware that they were being monitored and hence likely to be more compliant [12].
Differences in adherence rate reported could also be due to variations in the way it was defined, assessed and the methodology of the studies. The definition of compliance in this study was consistent with a few other studies which defined poor compliance as missing at least 1 injection in a week [4,10]. However, these studies were crosssectional whereas ours was a prospective study. Unfortunately, as there were limited suitable prospective studies which we could adopt, we decided to modify the model from Cutfiled et al. after consultation with our statistician. In our study, poor compliance was higher among GHD patients as compared to TS; however, this was not statistically significant possibly because the number of TS participants was small.
Our study showed poorer HV and HV SDS in participants with GHD with poor compliance (p< 0.05). This was consistent with the results published by Cutfield et. al [10] showing significantly greater linear growth in patients who had good compliance (≥85% compliance or missed less than 1 dose a week of injection). Similarly, Kapoor et. al [4] reported poor compliance (missing >1 dose per week) was associated with reduced height velocity (p<0.005).
The height velocity for children naïve to rGH is usually highest during the first year of treatment. Thus, we had done a separate statistical analysis excluding patients on the first year of rGH. The association of HV and HV SDS with compliance in GHD patients remained significant after this exclusion (Supplementary Table 1). However, no significant difference was found in TS patients, likely due to the small sample.
Two main factors were found to be significantly associated with poor compliance in this study, i.e., the duration of treatment and the age of the patients. There was a negative correlation between compliance and the duration of GH therapy. Participants with GHD who had poor compliance had a longer duration of treatment. There was also a significantly higher rate of poor compliance in the older age group of patients. These were consistent with the findings in other studies which reported poor adherence with longer duration of GH therapy [4,6] and among adolescents [3,14,23,24]. This further highlights the challenges in maintaining good compliance to treatment in chronic illnesses. Compliance in drug therapy is often poor in chronic non-lifethreatening conditions such as GHD [25]. Motivation may be low as the benefits are not immediately apparent and daily subcutaneous injections may present a significant burden [4,26]. Thus, regular checks on adherence and repeated motivation to patients are important throughout the course of GH treatment. This is especially important among adolescents given the unique developmental, psychosocial and lifestyle issues implicit in their transition to adulthood [27]. Defiance, rebellious, attention-seeking, and denial can be manifested by not taking required medications [24]. The process of transition from parental dependency to autonomy at this age also leads to confusion as to who is responsible for the administration of medication [14]. There were a few adolescents in our study who were studying at boarding schools away from home who cited lack of motivation and supervision as the reasons for poor compliance. Thus, extra attention and supportive treatment such as motivational and behavioral therapy may be needed to improve treatment adherence in adolescents. This is especially so for those who had been on therapy for a long duration of time.
There were no significant associations found between compliance with other variables which included demographic characteristics, gender, race, education level; personnel who administer the medication, supervision from caregivers, or types of devices. This was consistent with other studies which did not find a major influence of demographic factors on patient's adherence to medications [24]. However, one would expect involvement of parents and caregivers in drug administration to influence compliance [17,28].
Drug adherence in children is unique because of the involvement of a third party who is involved in supervision or administration of the medication, i.e., the parent/guardian [7]. Self-administration of medication had been reported to be associated with lower treatment adherence compared to administration by parents [14]. Our study did not find a significant association between supervision with compliance, possibly because of the small sample size. Our study showed there was only occasional supervision by 44% of parents/caretakers especially in the older age group (>10 years old). Parents or caregivers should be reminded to monitor administration regularly, even if the child or adolescent have been selfinjecting for a long period.
We also found no significant differences in compliance between the GH devices. This was similar to other studies [29] though some suggested improved compliance if patients were given the option to choose the device [4,22]. A few observational studies by Saizen ® using Easypod devices claimed high compliance rates can be achieved with precise and objective measurements of adherence [17,19,30]. However, caregivers or patients may have different preferences on the devices. The majority of our patients on GH treatment were using self-inject pen devices thus it would be difficult to compare the compliance rate between the two devices. Level of understanding is usually one of the main factors for poor compliance [31], and few studies suggested the association of the education level of caregivers with compliance [7]. This was not reflected in our study.
Reasons for missing medications reported in other studies include a short prescription of duration less than 4 weeks, scheduling issues (away from home), forgetfulness, inter-current illness and pain [4,23]. The majority of the participants in this study cited forgetfulness as the main reason for missing doses followed by insufficient medications, travelling away from home and being unwell. None of the participants had complaints of pain or needle phobia, which is a common belief of poor compliance [15,28]. All the participants were satisfied with their GH therapy and device used, though few of them were found to have technical issues in administration, which were promptly rectified. None complained of adverse effects. Recognition of reasons and factors contributing to poor compliance is necessary to enable remedial measures to be taken. Measures to address the common shortcoming of being forgetful include the suggestion of a reminder mechanism for example a calendar chart or reminder alarm, and commitment from parents to consistently supervise the injection. Provision of a longer supply of medication may also help with adherence, as inadequate supply of medications was a commonly cited reason for missing doses. In our current setting, most patients were required to collect new medications every 1-2 months at the hospital pharmacy thus logistics can be a major hinder.
Regular monitoring of adherence is important not just at initiation but throughout treatment as compliance tends to reduce with increased years of treatment. In our study, despite the frequent counseling given to the participants; missing injections still happen. We did not demonstrate a significant change in compliance rate over time in our study, likely because counseling was already provided at recruitment and the short duration of the study. However, we demonstrated that 8 (23.5%) participants with unsatisfactory GH administration techniques during the initial assessment improved at the end of the study.
Accurate methods for assessing medication compliance are also necessary. Ideally, patients should be instructed to bring their devices each visit to show proof of compliance and exchange used medication for every new collection of prescriptions. Doctors, pharmacists or specialized nurses can play an important influence on medication-taking behavior [24]. Regular counseling and review sessions may help reinforce drug compliance and identify cases at risk of poor treatment adherence. Patient knowledge including injection technique, transportation and storage of medication should be evaluated from time to time, as a longer duration of treatment is not necessarily associated with better techniques or understanding, as noted in our study.
The findings of this study had reaffirmed the importance of compliance to daily rGH injection to ensure a good treatment outcome. Education and creating awareness should be emphasized early from the time of initiating rGH, as patients and families may have a different perception of what constitutes poor compliance [7]. They may not realize missing an injection once a week could lead to a significant reduction in long-term growth velocity. More attention should be focus on poor compliance in adolescents and those with a longer duration of treatment, as they are at higher risk of poor drug adherence. Long-acting GH (LAGH) preparations are currently being developed [32,33]; however, these will not necessarily promise good drug adherence if the basic principles of education and constant monitoring were not practiced.
The strength of this study was the prospective nature enabling accuracy in clinical data, measurements of growth velocity and assessment of compliance. In addition to the calculation of returned vials and checking of devices, interviews based on a questionnaire to assess compliance were done to enable better assessment of adherence rates. At the same setting, reasons for poor compliance and confounding factors contributing to impaired growth velocity were also explored, such as issues with devices, the technique of drug administration, transport, and storage of the medication.
The limitation of this study was the small sample size, which could have resulted in the lack of significant associations for some of the variables analyzed, especially among patients with TS. Since this was not a blinded randomised study, the compliance and height velocity rate could have been artificially raised during the study. The interview sessions may hinder the participants from giving truthful answers and may explain the low incidence of needle phobia or dissatisfaction with the medication or devices demonstrated in this study.

Conclusion
Poor compliance resulted in a significantly reduced height velocity in patients with GHD and is more common in adolescents and those who had been on treatment for a longer duration. As forgetfulness is the commonest reason for missing an injection, parents and caregivers should play their role to regularly supervise the injections. Measures to improve compliance should address underlying reasons and monitoring adherence throughout therapy. This is important to ensure an optimal height outcome and prevent wastage of funds for public-funded treatment. Larger, multicentre double-blinded randomized studies in the future is needed to provide more evidence on the cause and effect of poor drug compliances.